Millions of people use platforms such as Google Maps to search for routes to their desired destinations. Recently, researchers and mapping platforms have shown growing interest in optimizing routes for criteria other than travel time, e.g. simplicity, safety, and beauty. However, despite the ubiquity of algorithmic routing and its potential to define how millions of people move around the world, very little is known about the externalities that arise when adopting these new optimization criteria, e.g. potential redistribution of traffic to certain neighborhoods and increased route complexity (with its associated risks). In this paper, we undertake the first controlled examination of these externalities, doing so across multiple mapping platforms, alternative optimizations, and cities. We find, for example, that scenic routing (i.e. “beauty”-optimized routing) would remove vehicles from highways, greatly increase traffic around parks, and, in certain cases, do the same for high-income areas. Our results also highlight that the interaction between routing criteria and urban structure is complex and effects vary from city to city, an important consideration for the growing literature on alternative routing strategies. Finally, to address the lack of open implementations of alternative routing algorithms and controlled routing evaluation frameworks, we are releasing our alternative routing and evaluation platform with this paper.
|Original language||English (US)|
|Journal||Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies|
|State||Published - 2017|